AGENT STUDIO OPEN
№ 01
Local-first IDE for MCP and A2A agents.
Connect, inspect, record, replay, compare model behavior, ingest OTEL spans, and export regression suites to CI.
STACK
MIT · Desktop · Browser · CLI · VS Code
AURAONE OPEN · LOCAL FIRST · MIT
Open source. Local first.
Four IDEs and a library of engines for agent reliability, rubric authoring, robotics review, and dataset trust. Public source. No account. Your data stays on disk.
REPOS
30+ public
Engines, IDEs, actions, packets — every one public, every one MIT.
TOOLS
4 local IDEs
Agent Studio, Rubric Studio, Robotics Studio, and the Open v2 set.
ADOPTERS
Labs · platforms · CI
Frontier labs, agent platforms, and PR review workflows already in.
FOUR SURFACES
Pick the IDE for the work.
Each one runs on disk, ships MIT, and exports a portable artifact a reviewer can pick up without a hosted account.
OPEN PAGE →
RUBRIC STUDIO OPEN
№ 02
The IDE for the rubric.
Local, file-based, git-friendly authoring for criterion-level evaluations. Author, test, calibrate, diff, and export.
STACK
MIT · Desktop · Browser · CLI
ROBOTICS STUDIO OPEN
№ 03
The IDE for reviewed teleop and VLA datasets.
Open LeRobot, RLDS, OpenX, HDF5, ROS bag, and mp4/jsonl captures. Scrub, tag, cluster, probe, export.
STACK
MIT · Desktop · CLI
OPEN V2
№ 04
Trust gates for agentic and embodied AI.
Twelve installable packages for MCP/A2A review, trace replay, robotics data quality, VLA diagnostics, and review packets.
STACK
MIT · pip · GitHub Actions
PUBLIC ENGINES
Read the source. Install in a venv.
The libraries underneath every Open IDE. Each runs from a notebook, a CLI, or a GitHub Action.
REPOPURPOSEINSTALLSOURCE
mcp-risk-linter
Risk taxonomy and lint pass for MCP server manifests.
pip install mcp-risk-linter
GITHUB →a2a-contract-test
Offline contract tests for A2A agent cards and task lifecycles.
pip install a2a-contract-test
GITHUB →tool-call-replay
Deterministic replay harness for failed agent tool calls.
pip install tool-call-replay
GITHUB →agent-trace-card
Portable Markdown and JSON card for one agent run.
pip install agent-trace-card
GITHUB →otel-eval-bridge
Bridge OTEL and Phoenix GenAI spans into eval regression cases.
pip install otel-eval-bridge
GITHUB →AURAONE OPEN
Your work. Your data. Your tools.
Read the source. Run it locally. Bring AuraOne in when shared state is the actual problem.